Lightroom settings: Prophoto RGB?

urbanaries

When I open a .Cr2 file in Lightroom for editing in CS2, it converts it to a TIF, and then opens the file.

But when I get the file into CS2, it says the color profile is mis-matched -- Prophoto RGB.

I am trying very hard to keep everything sRGB across the board, but I don't know how this is affecting the images, and if so, how to change it in LR 1.1.

When I export images, I always export in sRGB.

colourbox

It probably isn't going to hurt the images, as ProPhoto RGB is much larger than sRGB. If you want to stay in sRGB the whole way, then instead of using the Lightroom shortcut that always sends it to Photoshop as ProPhoto RGB (which works with my workflow), you'll probably want to use the Export command instead and customize the output color space to sRGB.

The way the direct link to Photoshop is set up, it looks like they wanted that particular route to lose as little info as possible. Because I think it isn't just ProPhoto RGB, but also 16 bits per channel, isn't it?

Andy

urbanaries wrote:

I am trying very hard to keep everything sRGB across the board, but I don't know how this is affecting the images, and if so, how to change it in LR 1.1.

urbanaries

colourbox wrote:

It probably isn't going to hurt the images, as ProPhoto RGB is much larger than sRGB. If you want to stay in sRGB the whole way, then instead of using the Lightroom shortcut that always sends it to Photoshop as ProPhoto RGB (which works with my workflow), you'll probably want to use the Export command instead and customize the output color space to sRGB.

The way the direct link to Photoshop is set up, it looks like they wanted that particular route to lose as little info as possible. Because I think it isn't just ProPhoto RGB, but also 16 bits per channel, isn't it?

I appreciate that sRGB is the smallest range of colors, but I'm printing primarily with SmugMug for now, and they require sRGB only.

Editing in a 16-bit space in CS2 isn't going to do me any good if I turn around and export in 8-bit space, is it? Or is it better to do pixel edits in a larger bit if possible?

urbanaries

Andy wrote:

thanks andy, helpful as always!

colourbox

urbanaries wrote:

I appreciate that sRGB is the smallest range of colors, but I'm printing primarily with SmugMug for now, and they require sRGB only.

That's fine, no quibbles there. Anyway, Andy has once again shown the way, you can disregard my "advice.":D

urbanaries

colourbox wrote:

That's fine, no quibbles there. Anyway, Andy has once again shown the way, you can disregard my "advice.":D

oh no, I appreciate your reply, and when I get the calibration color issues down with sRGB, I'm sure I will then have outgrown it's limitations and ready to tackle the next level. Maybe even print on my own someday!

arodney

EVERYTHING you're doing in LR is processed in ProPhoto RGB (linear gamma). Make an export preset for sRGB for those who just have to have it. You can export the data in ProPhoto for the demanding output needs.

urbanaries

arodney wrote:

EVERYTHING you're doing in LR is processed in ProPhoto RGB (linear gamma). Make an export preset for sRGB for those who just have to have it. You can export the data in ProPhoto for the demanding output needs.

Thank you for the helpful tip, I appreciate it!!! I'm starting to print more elsewhere so I am thankful for the help getting the best output.

NavyMoose

arodney wrote:

EVERYTHING you're doing in LR is processed in ProPhoto RGB (linear gamma). Make an export preset for sRGB for those who just have to have it. You can export the data in ProPhoto for the demanding output needs.

If I process in ProPhoto RBG and export the JPEG for SmugMug in sRGB, will the JPEG image quality suffer because the much larger color space I was working in was compressed to a much smaller workspace?

Thank you!

NavyMoose

arodney

NavyMoose wrote:

If I process in ProPhoto RBG and export the JPEG for SmugMug in sRGB, will the JPEG image quality suffer because the much larger color space I was working in was compressed to a much smaller workspace?

No. You will reduce the gamut (so there's clipping) but that's a fact of life in all color space conversions.

NavyMoose

I hate to sound stupid, but I have to ask what will probably be a dumb question.

I'm relatively new to digital and I'm learning all I can. Will the clipping cause the colors to go out of whack to the point a customer might be disappointed in the colors?

I know SmugMug has auto and true color in the ordering process. Is it possible the "auto" setting can solve the clipping problem and still produce an acceptable image?

Thank you

arodney wrote:

No. You will reduce the gamut (so there's clipping) but that's a fact of life in all color space conversions.

arodney

NavyMoose wrote:

Will the clipping cause the colors to go out of whack to the point a customer might be disappointed in the colors?

That should not happen with good profiles and color management. There should be no color shifts. A very saturated red just becomes less saturated. Of course, if you're dealing with printer profiles, YMMV depending on how the profile was built. But for color space conversions, no real issues.

You can test this yourself. Convert from a wide gamut space (ProPhoto RGB) to sRGB and make a print of both.

Tanuki

urbanaries wrote:

I appreciate that sRGB is the smallest range of colors, but I'm printing primarily with SmugMug for now, and they require sRGB only.

Editing in a 16-bit space in CS2 isn't going to do me any good if I turn around and export in 8-bit space, is it? Or is it better to do pixel edits in a larger bit if possible?

If you are only doing a lot of processing (not just minor tweaks) and you are starting with a RAW file, then you may see a benefit of working in a larger color space because you may need some of the color information that would have been clipped if you converted to sRGB before editing.

On the other hand, it may not make much practical difference for most cases. There are two types of color space conversions, one preserves colors (and can clip) while the other pulls colors into the gamut (but changes colors). If you don't know which one you are doing, then chances are that you are pulling colors into gamut. This should give decent results, even if it's not quite as good as working in the larger gamut.

I think you will be fine to work in 8 bit sRGB unless you don't like the results, in which case you can do that particular photo in 16 bit mode and a wide gamut color space.

Mike

arodney

On the other hand, it may not make much practical difference for most cases. There are two types of color space conversions, one preserves colors (and can clip) while the other pulls colors into the gamut (but changes colors). If you don't know which one you are doing, then chances are that you are pulling colors into gamut. This should give decent results, even if it's not quite as good as working in the larger gamut.

Moving from larger to smaller color spaces always reduces the gamut. You have two options sometimes! One is a colorimetric intent which clips out of gamut colors to the boundaries of the new color space. Those colors clip to the boundaries, period. Perceptual rendering maps all colors in a non uniform way so the appearance is maintained as best as possible. That's why the term perceptual is used. Its supposed to (perceptually) maintain the color relationship even if non out of gamut colors need to be affected. But the net result in both is, you toss away usable colors depending on the resulting color space gamut.

When you convert from working space to working space (ProPhoto RGB to sRGB as an example), there's only one option: Colorimetric. There's no perceptual table in simple matrix like profiles here. You're always clipping to the boundaries of the new space.

I think you will be fine to work in 8 bit sRGB unless you don't like the results, in which case you can do that particular photo in 16 bit mode and a wide gamut color space.

All conversions and editing of pixels (not Raw conversions) result in some data loss due to rounding errors. The problem with 8-bit workflows is you never know when, or how many subsequent edits will introduce posterization on output. Not at all an issue with high bit (16-bit as referred to in Photoshop although its really 15 bit) workflows. Sending the best 8-bits to the printer is the goal although more drivers are allowing users to send more data (Epson just started this, Canon has been doing it for awhile). When you only work with 8-bits, you just don't know when it will fall apart due to edits you may make today or in the future:

http://staging.digitalphotopro.com/tech/the-bit-depth-decision.html

Tanuki

arodney wrote:

Moving from larger to smaller color spaces always reduces the gamut. You have two options sometimes! One is a colorimetric intent which clips out of gamut colors to the boundaries of the new color space. Those colors clip to the boundaries, period. Perceptual rendering maps all colors in a non uniform way so the appearance is maintained as best as possible. That's why the term perceptual is used. Its supposed to (perceptually) maintain the color relationship even if non out of gamut colors need to be affected. But the net result in both is, you toss away usable colors depending on the resulting color space gamut.

When you convert from working space to working space (ProPhoto RGB to sRGB as an example), there's only one option: Colorimetric. There's no perceptual table in simple matrix like profiles here. You're always clipping to the boundaries of the new space.

Thanks for clarifying. I'm working within the limits of PSE, and was under the false assumption that the full version of PS had the ability to do perceptual rendering for color space transformations.

arodney wrote:

All conversions and editing of pixels (not Raw conversions) result in some data loss due to rounding errors. The problem with 8-bit workflows is you never know when, or how many subsequent edits will introduce posterization on output. Not at all an issue with high bit (16-bit as referred to in Photoshop although its really 15 bit) workflows. Sending the best 8-bits to the printer is the goal although more drivers are allowing users to send more data (Epson just started this, Canon has been doing it for awhile). When you only work with 8-bits, you just don't know when it will fall apart due to edits you may make today or in the future:

http://staging.digitalphotopro.com/tech/the-bit-depth-decision.html

In general, I couldn't agree with you more for someone doing serious post-processing and/or using their own printer, but didn't the original poster ask about Lightroom and Smugmug workflow? Lightroom works in 16 bit Prophoto space, but as far as I understand, Lightroom shows clipping for sRGB color space. So if Photoshop editing is only minor tweaks (dodge and burn, local contrast adjustment, etc) then it is unlikely to go out of gamut. And in regards to 8 bit versus 16 bit in Photoshop, rounding errors introduced by 8 bit space should be unnoticeable in the final print for minor tweaks.

Mike

arodney

Tanuki wrote:

Thanks for clarifying. I'm working within the limits of PSE, and was under the false assumption that the full version of PS had the ability to do perceptual rendering for color space transformations.

NOT with simple matrix profiles (working spaces). The tables do not exist. The options for Perceptual and Saturation are there, they do nothing (try it). The only intent you can produce is colorimetric.

And in regards to 8 bit versus 16 bit in Photoshop, rounding errors introduced by 8 bit space should be unnoticeable in the final print for minor tweaks.

It depends! Do you know for a fact how many edits you'll make today or in the future and when banding will show up due to such edits? Do you know what output devices you'll use in the future and how their dither might revel the data loss? With high bit editing, that's never an issue. With 8-bit editing, it's always looming as a possibility.

Tanuki

arodney wrote:

NOT with simple matrix profiles (working spaces). The tables do not exist. The options for Perceptual and Saturation are there, they do nothing (try it). The only intent you can produce is colorimetric.

I already got this point. You'll notice I had said that I was under a false assumption about PS full version capabilities.

arodney wrote:

It depends! Do you know for a fact how many edits you'll make today or in the future and when banding will show up due to such edits? Do you know what output devices you'll use in the future and how their dither might revel the data loss? With high bit editing, that's never an issue. With 8-bit editing, it's always looming as a possibility.

Good point... but if you decide to go back and do more significant edits, why not start over with 16 bit and a wider gamut at that time? That seems like a better way to go then to do all PP in 16 bit which, from what I understand, means that your PSD files will be so huge that you can't afford the hard drive space to save them all and therefore you will have to start over for your future edits anyway.

Mike

arodney

HD space is cheap, my time isn't.

I could go back BUT if I did a lot of work on the rendered image, going back and building a new Raw file in high bit, wide gamut means I start all over again. At least with a Raw workflow, you can do this but its still more time to do redundant edits when all I have to do is render the high bit, wide gamut pixel based file once.

NavyMoose

I shoot exclusively in RAW. I only do relatively minor tweaks in LR and rarely do I need to open CS3. I mainly make minor adjustments to exposure and if it is required, do the conversion to black and white.

Color management is something I learning about. Yesterday, I ordered Margulis' "Professional Photoshop, 5th Edition". I use Scott Kelby's book for how to do something in CS3 / LR. I calibrate my monitor using Spyder. I am not using any profiles so to speak. I am not using the ICC profiles for EZ Print or anyone else. I have the ones for MPix and I think WHCC but I am hesitant to use them until I learn more about how they work and interact with the whole digital work flow.

Thank you for your help!

arodney wrote:

That should not happen with good profiles and color management. There should be no color shifts. A very saturated red just becomes less saturated. Of course, if you're dealing with printer profiles, YMMV depending on how the profile was built. But for color space conversions, no real issues.

You can test this yourself. Convert from a wide gamut space (ProPhoto RGB) to sRGB and make a print of both.

arodney

NavyMoose wrote:

I calibrate my monitor using Spyder. I am not using any profiles so to speak.

Actually you are! There's the display profile. There's the document profile. Both are necessary to produce consistent and correct color appearance. Digital images are just big piles of numbers and numbers alone don't describe color appearance anymore than a recipe of ingredients without units of measurement don't provide a road map to make the dish. You may not be using output profiles but you're using profiles all the time in Photoshop and Lightroom. When you tell either you want sRGB or ProPhoto RGB, you're defining the scale of the RGB numbers using profiles. That's all profiles do, they define the scale of numbers for images or devices.

Output profiles are useful when you want to preview on screen a simulation (soft proof) of what the document numbers will look like when you print them.